The defect complexes that are formed when protons with energies in the MeV-range were implanted into highpurity silicon were investigated. After implantation, the samples were annealed at 400 °C or 450 °C for times ranging between 15 minutes and 30 hours. The resistivity of the samples was then analyzed by Spreading Resistance Profiling (SRP). The resistivity shows minima where there is a high carrier concentration and it is possible to extract the carrier concentration from the resistivity data. Initially, there is a large peak in the carrier concentration at the implantation depth where most of the hydrogen is concentrated. For longer anneals, this peak widens as the hydrogen diffuses away from the implantation depth. Following the changes in resistivity as a function of annealing time allows us to characterize the diffusion of hydrogen through these samples. Differences in the diffusion were observed depending on whether the silicon was grown by the magnetic Czochralski (m:Cz) method or the Float zone (Fz) method.

Electron beam induced current (EBIC) measurements were used to determine the doping type and to extract the diffusion length in proton implanted silicon wafers. This method makes it possible to distinguish between n-type and p-type at low carrier concentrations. Because of the defects caused by the implantation, the diffusion length is much smaller in the implanted than in the non-implanted regions.

Protons with energies of 1 MeV and 2.5 MeV were implanted into a p-doped silicon wafer and then the wafer was annealed at 350 °C for one hour. This resulted in two n-doped layers in the otherwise p-doped sample. The carrier concentration was measured using spreading resistance profiling while the positions of the four pn-junctions were measured using electron beam induced current measurements. The carrier concentration is not limited by the available hydrogen but by the concentration of suitable radiation induced defects.

Electron beam induced current (EBIC) measurements were used to produce cross sectional images of superjunctions in CoolMOS™ power transistors. The positions of the pn-junctions were determined by EBIC measurements. Knowing the exact locations of the pn-junctions is important for CoolMOS™ since it relies on the principle of charge compensation. For charge compensation, the donors in the n-doped regions must be compensated by an equal amount of acceptors in the p-doped regions. We show that EBIC can provide valuable input for process tuning and process simulations. This will enable the use of smaller dimensions and higher doping levels resulting in a lower on-state resistance . Superjunction transistorsConventional MOSFET (Metal Oxide Semiconductor Field-Effect Transistors) devices exhibit high on-state losses with blocking voltages V br exceeding 100 V. These on-state losses are caused by relatively high on-resistances in conventionally designed MOSFETs with blocking voltages higher than several hundred volts. The contribution of the lightly doped drift region to the on-resistance R DS,on increases superlinearly with V br . So-called superjunction MOSFETs, like Infineon's CoolMOS™, apply the concept of charge compensation, which allows significantly higher substrate doping of the drift region while at the same time maintaining a high breakdown voltage (1,2). In such devices, R DS,on only increases proportionally to V br 1.3 . By this, superjunction devices break the physical limit of the standard Si-MOSFET technology with respect to the area-specific turn-on-resistance value R DS,on ×A ( Figure 1). Hence, superjunction transistors are particularly attractive for applications calling for high blocking voltages in the range of 400 V to 1000 V. 10.1149/04901.0475ecst ©The Electrochemical Society ECS Transactions, 49 (1) 475-481 (2012) 475 ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 142.103.160.110 Downloaded on 2015-06-25 to IP